Rivers Piddle and Frome, and Poole Harbour

Alan Walker, Mike Pawson
Case Study Description
Rivers Piddle and Frome,
and Poole Harbour
Centre for Environment, Fisheries and Aquaculture Science (Cefas)
Lowestoft Laboratory
Pakefield Road, Lowestoft, Suffolk NR33 0HT.
Tel + 44 1502 524351, Fax + 44 (0) 1502 526351,
Mike.Pawson@cefas.co.uk; Alan.Walker@cefas.co.uk
1 Introduction
The River Frome is a chalk stream in SW England, with a main stem of 48 km in length. It
rises in the Dorset Downs at Evershot, and passes through Maiden Newton, Dorchester, West
Stafford and Woodsford. A key tributary of the River Frome is Tadnoll Brook, which enters
the river 10 km upstream of the tidal influence (Figure 1). The River Piddle rises in Alton
Pancras and flows for ~30 km south and then south-east, approximately parallel to the River
Frome. Two key tributaries of the River Piddle are the Bere Stream and Devils Brook. The
combined Frome and Piddle catchment area is 454 km².
Both rivers drain into Poole Harbour, which has an area of 38 km². The harbour is extremely
shallow (average depth: 48 cm), with one main dredged channel, and is characterized by
extensive mud flat and salt marsh habitats, as well muddy shores, sandy shores and seagrass
meadows.
There are no significant barriers to eel migration on either the Frome or the Piddle, other than
possibly the flood alleviation scheme in the upper Piddle (see below). The surrounding land
usage is predominately agricultural with no industry or major settlements within their
catchments. As a consequence, water quality remains very good and both rivers have the fauna
and flora typical of lowland chalk streams.
A flood alleviation scheme was completed in October 2005 at Piddletrenthide (2.5 km
downstream from the source of the River Piddle). Floodwater bypasses the village via a
culvert 560 m in length. Eels were present in this area in 2003, both above (density 0.02
eels.m -2 , n = 7) and below the village (density 0.14 eels.m -2 , n = 58). There is no information
regarding what, if any, effect the culvert may have on the eel population in this part of the
upper catchment. The rivers Piddle and Frome have been extensively studied for their trout
and salmon populations, fish habitats and invertebrates. Angling on both rivers is very
popular, especially for salmon, trout and grayling.
There are two silver eel racks, one on the River Piddle and one on the Frome. The Frome rack
tends only to be run for research purposes. The River Piddle rack operates each year during
the silver eel run and the catch is sold for UK consumption. This trap was fitted with new rack
in 2000 with 10 mm bar spacings to ensure all male silver eels were caught.
This document is part of the Report of FP6-project FP6-022488 ‘Restoration of the European eel
population; pilot studies for a scientific framework in support of sustainable management.’
lime - Study Leading to Informed Management of Eels
Sixth Framework Programme, Priority 8.1, Policy-oriented research, Scientific Support to Policies.

Poole Harbour supports a fyke net fishery. Local eel fishermen claim that stocks and fisheries
have declined and the average size of eels in catches has fallen over the last 20-30 years. The
situation was investigated for the Environment Agency (Knights & White, 1997) and a
summary of key points and conclusions is given below.
Fyke nets were introduced into the Harbour fishery in the 1950s and catches in the 1960-70s
were about 30 t year -1 (Morrice, 1989). At the end of the 1960s, five ‘punts' were laying up to
650 fyke-net ends per night. Increasing effort (as much as a tenfold increase in numbers of
ends deployed per day) was needed to maintain catches in the early 1980s. Subsequent
declines in catches have, it is claimed, led to reduced numbers of fishermen and hence total
effort since the mid-1980s. Given the area of Poole Harbour as 38 km 2 , and extrapolating from
estimated total commercial yields, Knights & White (1997) calculated that catches could have
fallen from about 7.9 to 1.3 kg ha -1 year -1 in the between the 1960-70s and the 1990s, a
decline of 80-85%. Although firm evidence is lacking, growth-overfishing is implicated,
probably exacerbated by reduced recruitment.
In relation to more recent changes, quantitative information on fishery yields was first
collected by the former NRA in 1988, licensing of the fishery began in 1991, and catch returns
are probably inaccurate. In 1988, the declared catch was 17,045 kg (including 1270 kg caught
by a tidal seine net), but there are no records for fishing effort. Between 1991-96, the average
catch was 19.6 (range 10.9-28.0) kg fyke-net-end -1 year -1 , shared between 2 to 8 licensees,
with no apparent trend.
Figure 1 The River Piddle, River Frome and Poole Harbour.
2 Fishing capacity
There is no limit on fishing capacity
p.340 lime FP6-project 022488

3 Fishing effort
There is no limit on fishing effort, the total fishing effort recorded in 2005 was 800 fyke net
days.
4 Catches and Landings
In 2005 the declared catch was 300 kg of yellow eel and 126 kg of silver eel.
5 Catch per Unit of Effort
In 2005 the CPUE was 0.53 kg per fyke net per day
6 Scientific surveys of the stock
6.1 Recruitment surveys
No surveys on glass eel/elver recruitment have been carried out for the Piddle/Frome system.
6.2 Yellow eel surveys
Electric fishing surveys were conducted throughout the Piddle in 1977, 1999 (16 sites), 2003
(20) and 2004 (5). The data set for 1977, based on two single-pass runs of 14 km one month
apart, does not include eels

6.3 Silver eel surveys
Silver eel were measured for length and weight from catches at licensed fixed eel racks fished
on a varying number of nights over the period of the silver eel run from September to
December, in 1977 (Piddle) and 2003–2004 (Piddle and Frome).
7 Catch composition by age and length
‘Catch’ composition by age and length are available for all the electric fishing and fyke netting
(yellows), and eel rack (silvers) surveys.
8 Other biological sampling (age and growth, weight, sex,
maturity, fecundity).
See Section 6, above.
9 Literature references.
Knights B. & White E.M. (1997) An appraisal of stocking strategies for the European eel,
Anguilla anguilla L. In Stocking and Introduction of Fish (ed. I.G. Cowx), pp. 121-140.
Fishing News Books, Oxford.
Knights, B., Bark, A., Ball, M., Williams, F., Winter, E. and S. Dunn (2001). Eel and elver
stocks in England and Wales – status and management options. Environmental Agency,
Research and Development Technical Report W248. 294 pp.
Morrice, C.P. (1989) Eel fisheries in the United Kingdom. MAFF Internal Report No. 18,
33pp. MAFF, Lowestoft, England
10 Application of the Models
For the purposes of testing models in SLIME, it was decided to focus on the freshwater
portion of the Piddle catchment and exclude the estuarine Piddle Harbour, since the latter is
the site of a mixed stock eel fishery (with the River Frome).
10.1 SMEP
Two models were applied to the Piddle dataset; SMEP and GlobAng.
The Piddle population was appropriate for modelling with SMEP because catchment data
were readily available, along with eel population data sufficient to inform the parameters file.
10.1.1 Parameter Setting
At the time of the workshop, SMEP applied growth according to the Von Bertalanffy model,
and ‘UK eel’ default values for L∞ and k were used. It was recognised, however, that the VB
model is probably not appropriate for simulating growth of eel, and this is an area of continued
model development.
Recruitment: No direct recruitment measures were available for the Piddle, but a trend was
applied to simulate changes in CPUE of UK glass eels fisheries since 1980. Elvers recruit to
the Piddle after a period of growth in Poole Harbour: the mean length of recruits was
increased above the default value for glass eel in order to reflect this.
p.342 lime FP6-project 022488

Carrying Capacity (biomass): maximum densities observed for streams were converted to
biomass, although these may still be well below maximum possible biomass of yellow eel
standing stock. Both the concept, and representative values, of carrying capacity for eel
require further exploration.
10.1.2 Population Data
Preliminary estimates based on mark-recapture exercises suggest that recent silver eel
escapement from the Piddle is in the region of 6000 eels per year.
The most complete data set for yellow eel was for 2003, so this was used as a reference year,
against which to compare the model outputs in terms of the yellow eel production.
10.1.3 Model Outputs
Modelling with SMEP suggested that a recruitment of 1.2 million elvers would be required to
produce an average annual silver eel escapement of 6000 eels. Based on this level of presentday
recruitment, the SMEP model suggested that historic (pre-1980) recruitment and silver eel
escapement would have been 4.8 million elvers and 24,000 silver eels, respectively. However,
both modelled silver eel runs were dominated by male eels (97-100%), whereas catches of
silver eels since 1999 have had much higher proportions of females (46-99%). Furthermore,
SMEP required yellow eel densities in the three reaches that were orders of magnitude higher
than observed in order to produce this level of silver eel escapement.
Calibrating the model with the yellow eel densities observed in 2003 suggested a recruitment
of about 70,000 but a silver eel run of only 400+.
10.1.4 Conclusion
SMEP has definite potential for modelling eel populations and silver eel output, but will
require considerable further parameter testing and development to cope with application to
data-poor rivers. Describing the catchment in reaches, setting an appropriate level of
recruitment where no actual data exist, and using parameter settings suitable for the river type
appear be the key constraints to useful model outcome. These three aspects are all featured in
the ongoing development of SMEP.
10.2 GlobAng
GlobAng is designed to perform simulations of eel population dynamics within a
hydrographical network. After calibration with real field data, the model is able to evaluate
silver eel escapement in response to a variety of management scenarios, especially when
spatial dimension is important. The model integrates ageing, recruitment, sexual
differentiation, silvering, natural mortality and for the first time migration within a watershed.
Fishing mortality and impacts of migration barriers can also be simulated.
Similar data were applied to the GlobAng model as were applied to SMEP and the same
recruitment and carrying capacity comments made under SMEP apply here. However, as
GlobAng requires reaches of equal length, the three SMEP reaches were split into seven, each
of nearly 10 km long. Also, as GlobAng is an age-related model, the observed length
distributions were transformed to age distributions using a length-age key developed from the
2003 data.
10.2.1 Model Outputs
Calibration of the model with the recruitment trend and the observed yellow eel population in
the three reaches in 2003 suggested a historic (pre-1980) recruitment level of around 550,000
FP6-project 022488 lime p. 343

and an annual escapement of 6250 silver eels, most of which were female. According to the
CPUE trend, therefore, the present-day recruitment would be about 140,000 elvers, which the
model suggests should produce about 3000 silver eels per year, most of which will be females.
The model overestimated the abundance of young eels, especially in the lower reaches.
However, the model set-up required recruitment from Poole Harbour to reach 1 at elver age 2,
whereas, as suggested by PL and others before, Poole Harbour may be acting as a recruitment
sink from which elvers migrate upstream at a range of ages.
The model underestimated both the abundance of older eels throughout, and the average age
of silver eels, at 4 and 10 years for males and females, whereas the average age of silver eels
sampled in 2000 to 2004 was 17 years, and ranged from 9 to 29 years old. This suggests that
the growth and probably maturation modelling were inappropriate for the Piddle eel
population.
10.2.2 Conclusion
GlobAng predicted a level of present-day silver eel escapement that, while it was half that of
the mark-recapture estimate, was a much better prediction than that provided by SMEP.
Improvements are needed to simulate a more effective colonisation of the catchment, age
structure of silver eels and to shift the sex ratio in favour of males. It was noted, however, that
the observed densities did not correspond to the typical case of diffusive colonisation with
higher abundance in the reaches closest to the estuary. Perhaps, therefore, a more
comprehensive survey programme would facilitate better representation of the yellow eel
stocks in each reach, and allow GlobAng to better model the Piddle eel stock.
p.344 lime FP6-project 022488